��Դ�ͳ�Ķ༶��ά��ֵ�˲����ں��������ڴ�С�Ե������������������Ӱ�죬�����һ������Ӧ��Ȩ�Ľ����ڶ༶��ά��ֵ�˲������Դ�ͳ�Ķ༶��ά��ֵ�˲����ں������иĽ���ʹ����б������Ժ�ϸ���������ڸĽ����ں����Ļ����ϣ����������Ӧ��Ȩ��������Ȩ����������Ӧ�Զ�����˥������Ч�źŵı���춨�˻���������Ӧ��Ȩ�Ľ�������ֵ�˲���ȥ���������ԡ�ͨ������ģ�ͺ�ʵ�����ݴ���ĶԱȣ�����������ȥ�������ͱ�����Ч�ź��������ڴ�ͳ�Ķ�ά�༶��ֵ�˲�����

Window function and window size of traditional multilevel 2D median filter influence the noise processing of seismic data�� We propose an adaptive weighted improved window multilevel two-dimensional median filter�� The window function of traditional multilevel 2D median filter is improved��which protects linear and minutiae feature�� Based on the improvement of the window function��we propose an adaptive weighted function which lays a foundation for the noise attenuation and effective signal fidelity��so adaptive weighted improved window median filter can remove the noise significantly�� Through the comparison of the theoretical model to the actual data processing�� the results show that the method is superior to traditional 2D multilevel median filter in noise attenuation and effective signal protection��